Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
10 PRECISION SERVO CONTROL SYSTEM
FOR A PNEUMATIC ACTUATOR
CLAIM TO PRIORITY
The present invention claims priority to two co-pending United States
Provisional patent
applications, the first of which is application number 60/161,875, filed
October 27, 1999, entitled
"Pnuematic Servo System For A Linear Actuator," and the second of which is
application
number 60/240,765, filed October 17, 2000, entitled "Pnuematic Servo System,"
both of which
are hereby incorporated by reference.
FIELD OF THE INVENTION
The present invention relates generally to the field of control systems for
pneumatic
actuators. More specifically, the present invention relates to a servo control
system that provides
precise and repeatable control of a pneumatic actuator.
BACKGROUND OF THE INVENTION
Mechanical actuators are used in a variety of industrial applications to move
machine
elements from one position to another. There are three different ways to power
the movement of
an actuator, electrically, hydraulically or pneumatically. Electrically
powered actuators are used
in situations requiring precise control and repeatability. An electrically
powered actuator such as
a screw drive or belt drive system powered by a rotary servo motor system has
the ability to
move at different speeds and to stop at any location along the entire length
of the stroke of the
actuator. Unfortunately, electrically powered actuators are prohibitively
expensive for
applications involving moving large loads or moving loads at rapid speeds.
Hydraulically and
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
pneumatically powered actuators, on the other hand, use a fluid (oil for
hydraulics vs. air or gas
for pneumatics) to provide a substantial force to a piston that moves inside a
chamber and is
connected to the actuator. Consequently, hydraulic and pneumatic actuators can
move large
loads and can move those loads at rapid speeds if desired. Hydraulic and
pneumatic actuators
also tend to be more durable than electrical actuators.
While hydraulic actuators are well suited for many applications, their use is
limited to
those envirorunents where oil can be used as part of the machine. For many
applications, it is
often not practical to utilize oil as the fluid to power an actuator.
Pneumatic actuators also tend
to be less expensive than hydraulic actuators or electrical actuators. The
problem is that
pneumatic actuators are much more difficult to precisely control than
electrical actuators or even
hydraulic actuators. Consequently, most pneumatic actuators are designed to
position the
actuator at only two stop positions, one at each end of the stroke of the
actuator where the end of
the chamber, a stopper or the like serves to physically stop the travel of the
piston, thereby
positioning the actuator at one of these two stop positions. This kind of two
stop pneumatic
actuator is controlled simply by supplying pressurized air to one side of the
piston until the
piston reaches the end of the stroke.
To control a pneumatic actuator to stop at positions other than the ends of
the stroke of
the actuator, the most common technique is to supply air at different
pressures to both sides of
the piston. Initially, this differential pressure will start the piston moving
in a direction from the
side with the higher pressure to the side with the lower pressure. Once the
piston is moving, this
differential pressure is reversed to cause the piston to stop moving. Ideally,
the differential
pressures can be applied to cause the piston to start and stop exactly at any
desired location along
the stroke of the actuator. In reality, adjusting the differential pressure to
achieve the delicate
balance required to precisely control the stop positions of the actuator is
quite difficult due at
least in part to the compressibility of the air or gas that is used as the
fluid to power and control
the actuator. These problems are compounded in situations involving changing
loads, long
stroke lengths or vertically-oriented stroke directions, or in situations
where the pressure of the
air or gas used to power the actuator is not tightly controlled.
The most common way of adjusting the differential pressure for this kind of
pneumatic
actuator is by controlling a variable valve or pair of variable valves, which
are sometimes
referred to as proportional valves or servo valves. Examples of control
systems developed for
differential pneumatic actuators that use proportional valves to control the
differential pressure
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
are described in U.S. Patents Nos. 4,481,451, 4,666,374, 4,790,233, 4,819,543,
5,154,207 and
6,003,428 and German Patent DE 3313 623 A1. Other variations on controlling a
differential
pneumatic actuator are described in U.S. Patent No. 4,878,417 which varies the
proportional
flow of the fluid in response to measurements from an accelerometer and U.S.
Patent No.
5,424,941 which uses a control system that converts differential pressure into
a differential mass
flow of the air that moves the piston in an attempt to minimize the problems
caused by the
compressibility of air.
An alternative technique for controlling the differential pressure of a
pneumatic actuator
is to use pulse width modulation (e.g., different widths of control pulses) to
control the supply of
pressurized air to both sides of the piston. Instead of turning a servo valve
part way on to control
the rate that air flows through the valve, pulse width modulation controls the
rate by quickly
turning the valve on and then off such that the average time the valve is on
is equivalent to the
proportional setting of a valve turned part way on for the same period of
time. Examples of this
pulse width modulation technique are described in U.S. Patents Nos. 4,628,499,
4,763,560 and
4,907,493. U.S. Patent No. 4,741,247 describes a very slow version of a pulse
width modulation
scheme where a series of step volumes of air are introduced into the chamber
one at a time in
order to move the piston a distance equal to the step volume.
Various attempts have been made over the years to address the problems caused
by using
air as the fluid to power a pnuematic actuator. One approach has been to use
some form of a
brake to assist in stopping the piston or the actuator. German patent DE
2,327,387 describes an
early use of an electromagnetic friction brake to stop a pneumatic actuator.
This patent uses a
conventional proportional servo valve to control the differential pressure.
Once the actuator
passes by a predetermined starting point for braking, the electromagnetic
friction brake is applied
intermittently to slow the piston down until it is moving at a much slower
speed, at which time
the brake is applied continuously to completely stop the actuator.
One of the problems with using a brake, however, is that the brake surface
will wear
down with repeated use and this results in variability in how accurately the
system operates over
time. U.S. Patent No. 4,106,390 describes a pneumatic linear actuator where a
pneumatic
mechanical brake is activated in a braking cylinder separate from but
connected to the piston to
prevent wear directly on the piston. The pneumatic mechanical brake is only
applied to stop the
actuator after a three-stage series of air braking decelerations are performed
by operating
solenoids in response to output signals generated by a sequence generator.
3
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
Other types of brakes have also been used as part of a control system for a
pneumatic
actuator. U.S. Patent No. 4,932,311 describes a pneumatic actuator having a
magnetic rotary
brake coupled to the piston by a ball screw shaft. A two-stage braking scheme
based on a target
braking speed is used to control the stopping locations of the piston. Once
the piston passes the
location where braking has been programmed to start, either or both an air
braking arrangement
and the magnetic brake may be applied at different periods along a braking
process in either an
intermittent or a continuous mode to keep the speed of the piston on target
with a calculated
braking speed. By attempting to control the speed of the actuator to match the
calculated braking
speed, the patent seeks to regulate the pneumatic actuator in a way that can
tolerate and
compensate for changes in the system, including changes in the brake.
A recent example of a controllable pneumatic actuator that uses a rotary
proportional
magnetic brake is described in PCT Publ. No. WO 00/53936. Both a simple
control system and
a sophisticated control system are described. In the simple control system,
differential pressure
is applied though a three-position solenoid valve to drive the pneumatic
actuator until it is within
1 S a defined distance (Ox) on either side of the desired stopping position
that is referred to as a
tolerance band. Once inside this tolerance band, the three-position valve is
set to a neutral
position where no pressurized air is applied to either side of the piston and
the proportional
magnetic brake is used to stop the actuator. If the load cannot be stopped
within the tolerance
band it will overshoot the desired stopping position and a reverse thrust must
be applied to move
the load back into the tolerance band. The problems with the simple control
system are that with
large loads and speeds where the kinetic energy is high and with system
friction present, it
becomes difficult to deal with the compressed air required to overcome the
system friction and
start the load moving and then stop the load moving in time to keep it within
the tolerance band.
This latter problem is a very undesirable problem known as hunting where the
actuator goes back
and forth about the desired stopping position or in this case, the defined
tolerance band, before
finally being stopped.
The sophisticated control system described in this PCT application attempts to
solve the
problem of hunting in those cases where the kinetic energy of the load being
moved is greater
than the braking force of the magnetic brake. In this case, the control system
calculates a shut
down point that will be prior to the desired stopping point based upon the
kinetic energy of the
system and the available braking force. The shut down point represents the
exact time that the
system will shut down the differential pressure applied to the actuator and
start the magnetic
L/
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
brake. The control system allows a user to input a desired velocity profile
for the actuator and
uses a low level brake signal to maintain that velocity at the programmed
desired velocity up to
the shut down point. While the calculation of a shut down point based on
kinetic energy may be
helpful in reducing the hunting problem of the simple control system, it does
not address the real
world problems of precision and repeatability in accurately positioning the
actuator at a desired
stopping position. To be effective, the calculation of the shut down point
requires precise
knowledge about the mass of the load and the variability of the braking force.
Even if precise
knowledge of these values can be programmed into the control system, this
sophisticated control
system is unable to accommodate changes in loads during an operation or
changes in the
remaining parts of the system due to wear or variance.
In spite of these various attempts, the control and use of pneumatic actuators
has been
unable to match the precision and repeatability of electrically powered
actuators. The goal of
being able to program a pneumatic actuator in the same way that an electrical
actuator is
programmed is well known, but has yet to be achieved. Ideally, a pneumatic
actuator could be
controlled merely by setting a desired motion profile and/or stopping position
and then relying
on a servo control system to use positional feedback to guarantee that those
results will be
achieved. If this were possible, pneumatic actuators could be used in a
variety of situations that
up to now have been the exclusive domain of electrical actuators.
Unfortunately, the number of
variables that must be controlled in a pneumatic system, and especially the
variabilities caused
by the compressibility of air, have frustrated the many attempts to realize
this goal.
Because of the complexities and variables involved in operating a pneumatic
actuator,
most existing control systems are highly individualized to the particular
pneumatic actuator and
typically require the operator to program the actuator by programming control
values for the
system in terms of encoder values, pulses, sequences or other units that have
complicated
relationships to a desired position or velocity. For those control systems
which utilize servo
systems, the operator must program a very large number of gains in an attempt
to fine-tune the
control system for a given application. While it may be possible to fine-tune
the programming of
a given pneumatic actuator with this very large number of gains or adjustments
to perform
adequately under the known conditions for that given application, even the
existing control
systems for pneumatic actuators that use servo systems are unable to
compensate for situations
involving changing loads, long stroke lengths or vertically oriented stroke
directions. For this
reason, existing control systems for pneumatic actuators limit their
performance claims to certain
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
defined conditions, such as limits on the ability to change loads, and always
express tolerances in
terms of percentages of the stroke length as a way to hide repeatability
errors associated with the
variability of larger volumes of air that are involved in longer stroke
lengths.
It would be desirable to provide an easily programmable servo control system
for a
pneumatic actuator that could effectively control the operation of the
pneumatic actuator to
match the precision and repeatability of electrical actuators, even under
changing loads, long
stroke lengths or vertically oriented stroke directions and that did not
require tightly controlled
air supplies and expensive servo valves to accomplish this precise servo
control.
SUMMARY OF THE INVENTION
The present invention is a precision servo control system for a pneumatic
actuator that
has a piston positionable over a stroke of the pneumatic actuator using a
supply of pressurized
gas. A brake and a sensor system are operably connected to the actuator and to
the servo control
system. The servo control system operates to initiate the forward thrust from
the pressurized gas
to move the piston along the stroke. When the piston reaches a deceleration
point along the
stroke as preferably determined by a programmable motion profile, the servo
control system
initiates a reverse thrust from the pressurized gas while maintaining the
forward thrust and
simultaneously begins to selectively apply the brake to stop the piston within
a predetermined
tolerance of a desired stopping position. The precision servo control system
is easily
programmable in a manner similar to that of a control system for electrical
actuators with respect
to both the motion profile and the limited number of gains of the control
system. The precision
servo control system can maintain the predetermined tolerance even under
changing loads, long
stroke lengths or vertically oriented stroke directions. The precision control
servo system utilizes
simple two position valves, instead of complex and expensive servo valves, to
regulate the
pressurized gas. The precision servo control system achieves positional
repeatability to
predetermined tolerance that is a fixed value, regardless of the stroke
length.
Preferably, the precision servo control system allows a user to input a
desired motion
profile for the actuator and uses a low level brake signal to limit the
acceleration to the desired
acceleration and velocity of the programmed motion profile until the piston
reaches the
decelerationpoint. Preferably, the brake is a proportional magnetic brake that
produces
increasing braking torque with increasing current applied to the brake. At the
deceleration point,
the precision servo control system continues to apply gas pressure to the
forward side of the
6
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
piston and also applies an equal and opposite gas pressure to the backward
side of the piston.
Because these two pressures are equivalent and opposite, they operate to
effectively cancel out
any contribution of the pressurized gas in the system after the deceleration
point. The active
application of the same pressurized gas to both the forward and reverse sides
of the piston
guarantees that there will be no complicated interaction of a reduction of
pressures during the
deceleration period.
In addition, the precision servo control system also monitors for a
predetermined
condition after the deceleration point at which the reverse thrust of
pressurized gas applied to the
backward side of the piston will turned off. The predetermined condition
represents the point at
which the precision servo control system first determines that either a
defined percentage of a
distance from the deceleration point to the desired stopping position has been
reached or that
there has been a defined percentage reduction in the velocity of the piston.
The predetermined
condition is selected such that the momentum of the load carried by the piston
will be
sufficiently small enough that it can be accurately controlled by the brake.
Removing the reverse
thrust at this point insures that the actuator will have the necessary forward
momentum to arnve
at the desired position without any jerking or stopping of the actuator. In
the event that the
momentum of the load carned by the piston does cause an overshoot, the
precision servo control
system locks the brake, then removes the forward thrust and applies the
reverse thrust, after
which the brake is slowly released to return the piston to the desired
stopping position.
Unlike the existing control systems that rely purely on positioning the piston
by adjusting
the pressure of the column of air on each side of the piston through use of
servo control valves or
PWM techniques, the precision servo control system of the present invention
can use relatively
inexpensive directional valves. More importantly, there is no need for the
pressurized gas supply
to be closely regulated. This means that the precision servo control system of
the present
invention can tolerate contaminated gas lines and other changes over time in
the pressurized gas
supply system because the same pressurized gas is supplied to both sides of
the piston during the
deceleration period. Because of this equal and opposite application of
pressure, the
characteristics of the gas supply on each side of the piston cancel each other
out. As a result, the
servo precision control system of the present invention does not need to
create complicated
higher orders control functions to compensate for the very complicated
characteristics of a
compressible and changeable gas supply.
WO 01/31205 CA 02388576 2002-04-25 PCT/LJS00/29877
In a preferred embodiment, the precision servo control system is programmed
utilizing
the standard control parameters of a servo-system including: (1) proportional
gain, KP; (2)
integral gain, KI; and (3) derivative gain, KV. The proportional gain, KP, is
the position error
gain that determines how sensitively the control system will respond to the
position error
(difference in command and actual position). The integral gain, KI, is the
position error integral
gain that determines how the control system responds to accumulated position
error while the
piston is approaching the target position. The derivative gain, KV, is the
speed error gain that
determines how effectively the controller/drive 100 responds to the speed
error while the piston
is in motion. All three of these control parameters are standard control loop
gains for a servo
system for an electrically powered actuator. In addition to these standard
control loop gains, the
precision servo control system preferably uses a deceleration compensation
path that enters into
effect after the braking point. The deceleration compensation path uses a
control parameter
beyond the standard control loop gain parameters of a servo system for an
electrically powered
actuator. This control parameter is a deceleration current constant gain, KT
that is used to set the
1 S minimum brake current while decelerating the load carried by the actuator.
The KT gain is used
to adjust for position overshoot, position undershoot, and deceleration
profile linearity when the
actuator approaches the target position. Unlike the complex and higher order
control schemes
that have been attempted to accommodate for all of the variables in a
pneumatic actuator system,
the present invention is capable of accurately and repeatably controlling the
positioning of a
pneumatic actuator with just these four gain parameters KP, KI, KV, and KT.
BRIEF DESCRIPTION OF THE DRAWINGS
Figures 1-4 are graphs showing the operation of various prior art control
systems for
pneumatic actuators.
Figure 5 is a graph showing the operation of a control system for a pneumatic
actuator
that utilizes a proportional magnetic brake.
Figure 6 is a graph showing the operation of a precision servo control system
for a
pneumatic actuator in accordance with the present invention.
Figure 7 is a side view, with portions removed, of a pneumatic position and
velocity
control system as applied to a linear actuator, i.e., a rodless cylinder, in
accordance with the
present invention.
Figure 8 is an end view as viewed along the line 8-8 of Figure 7 of the
rodless cylinder
WO 01/31205 CA 02388576 2002-04-25 PCT/US00/29877
of Figure 7 showing one of the sealing heads and its relationship to portions
of the cylinder and
positioning means.
Figure 9 is an exploded isometric view of a system for implementing an
embodiment
consistent with the present invention, including portions of a rodless
cylinder and associated
structure and a computer for controlling operation of the rodless cylinder.
Figure 10 is an isometric view, cut in half, of the positioning assembly for a
rodless
cylinder in accordance with the present invention.
Figure 11 is an isometric view of an assembled assembly of Figure 4 showing a
portion
of the cylinder.
Figure 12 depicts a first valve configuration that may be used with the system
of the
present invention.
Figure 13 depicts a second valve configuration that may be used with the
system of the
present invention.
Figure 14 is a block diagram of a controller/drive of the present invention.
Figure 1 S is a facial configuration of the controller/drive of the present
invention.
Figure 16 depicts the system of the present invention wherein a linear encoder
is utilized.
Figure 17 depicts the system of the present invention wherein a linear encoder
and a
friction pad braking device is utilized.
Figure 18 depicts the system of the present invention wherein a slide rod
actuator is
utilized.
Figure 19 depicts the system of the present invention wherein a rotary
actuator is utilized.
Figure 20 comprises various graphical examples of motion profiles that may be
used by
the system of the present invention.
Figure 21 is a flow chart of the operation of the system of the present
invention.
Figure 22 is a block diagram of the control scheme of the present invention.
Figure 23 is a motion profile including system operational indicators.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
To better understand the present invention, various ways in which control
systems have
been used to control the position of a pneumatic actuator will be described
and then compared to
the way in which the precision servo control system of the present invention
provides for precise
control and repeatability of positioning a pneumatic actuator. The graphs
presented in Figures 1-
CA 02388576 2002-04-25
WO 01/31205 PCT/LTS00/29877
6 are presented for the purpose of demonstrating the various concepts embodied
by the present
invention as compared to other control systems for pneumatic actuators and are
not graphs of
actual physical data collected from these systems.
Figure 1 is a graph showing the relation over time (shown on the x-axis)
between the
velocity (V) of the piston in a pneumatic actuator, the position (P) of the
piston and the idealized
thrust (T) of differential air pressure in accordance with a simple prior art
control system to
control the position of the actuator. The thrust (T) is applied in a forward
direction (f) at point
Xo to move the piston in a forward direction. At point X(t;) the differential
pressure is changed
and the thrust (T) is applied in a reverse direction (r) to stop the piston at
point X(tf). If there
were no variables in this system, the point X(t;) could be easily calculated
and the ability to
control the actuator to stop at point X(tf) would be simple. Unfortunately,
there are a large
number of variables that can affect the system and the chances are very small
that the actuator
will actually stop at point X(tf). The challenge of getting the actuator to
stop at or near the same
desired stopping point when the process is repeated is even more difficult.
The theoretical or
1 S ideal curve that the piston should follow in this system is shown in the
solid line. The dotted
lines show how the piston is more likely to behave in real world conditions.
It will be seen that
the type of simple control system shown in Figure 1 is not a true servo
control system because
there is no process to feed information about how the piston is moving back
into the process of
controlling the differential pressure. The only control provided by this
simple control system is
to change the location of point X(t;) where the thrust of the differential
pressure is reversed.
Figure 2 is a graph similar to Figure 1 that shows a conventional servo
control system
that uses a servo valve or pulse width modulation (PWM) technique to change
the thrust (T) in
response to sensor information about how the piston is moving. Like the simple
control system
of Figure 1, there is a point X; where the orientation of the thrust is
reversed. Unlike Figure l,
position and/or velocity information is used by the control system to adjust
the amount of
forward or reverse thrust (T) that is applied. For example, once the velocity
of the piston reaches
a defined value (V,), the control system can adjust the PWM signal at point
X(ta) to reduce the
thrust in an attempt to maintain the velocity at the defined value (V1). A
similar situation
happens with a servo control valve, only the value of the thrust (T) would be
controlled by
altering the amount of positive thrust delivered via the servo valve. In a
similar manner, the
control system can control how much reverse thrust is applied to stop the
piston. The control
system will adjust the PWM signal as point X(tb) to try to slow the piston
down so that it will
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
stop at the desired location X(tf). While this process has a better chance of
stopping the piston
near the desired location X(tf) than the control system of Figure l, the
variables due to the
compressibility of air, system friction, system wear, air pressure changes,
changing loads, long
stroke lengths or vertically oriented stroke directions are simply too much
for this servo control
system to be able to control. Again, the dashed lines show how the actuator of
this system is
likely to behave under real world conditions.
Figure 3 is a graph similar to Figure 1 but also showing a brake (B) applied
to stop the
piston after using the three-stage air braking decelarations as described in
U.S. Patent No.
4,106,390. In this system, a sequence generator outputs different control
signals to control a
series of solenoids according to a programmed sequence. First, a solenoid is
opened to provide a
forward air pressure to the piston to start the piston moving at point X(to).
At the same time, the
reverse side of the piston is opened to the atmosphere through a controlled
relief valve to create a
constant differential pressure that moves the piston at a constant low
velocity. When the piston
advances a programmed number of counts beyond a sensor at location point X(t,)
the sequence
generator changes the solenoid to create a first deceleration that occurs
during a time period (t)
when air pressure is applied to both the forward side (top line) and reverse
side (bottom line) of
the piston. After the time period (t) which can expire at any number of
different positions X(t2)
depending upon the velocity (V) originally achieved by the system, the
sequence generator again
changes the solenoid to effect a second deceleration where the solenoid is
changed to provide a
reverse air pressure and the forward side of the piston is opened to the
atmosphere. When the
piston advances a fizrther distance X(3), a third deceleration to a very low
velocity is
accomplished by slowing opening the reverse side of the piston to the
atmosphere. Finally, after
a final movement at the very low velocity to the target position, another
solenoid is opened to
apply the pneumatic mechanical brake (B) to the brake cylinder to stop the
actuator at location
X(tf). This control system provides a greater accuracy than the other control
systems, but does so
at the significant cost of dramatically increased time for movement of the
piston due to the very
slow velocities used in the final stages of approaching the desired location.
Even with this
system, there would be variances as reflected by the dashed lines that do not
allow for precise or
repeatable control of the piston.
Figure 4 is a graph similar to Figure 3 showing the control system and brake
arrangement
as described in U.S. Patent No. 4,932,311. A two-stage braking scheme based on
a target
braking speed is used to control the stopping locations of the piston. An
encoder on the ball
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
screw shaft determines the position of the piston, from which a speed of the
piston is calculated.
To start movement of the actuator at point X(to), air pressure is supplied to
a forward side of the
piston and the magnetic brake (B) is released. Once the piston passes the
location X(t;) where
braking has been programmed to start, either or both an air braking
arrangement and the
magnetic brake may be applied at different periods along a braking process in
either an
intermittent or a continuous mode to keep the speed of the piston on target
with a calculated
braking speed. The continuous mode of the air braking arrangement uses a
conventional simple
differential pressure technique where forward pressure is turned off and
reverse pressure is
applied. The intermittent mode of the air braking arrangement uses a variation
on the pulse
width modulation control technique to create a differential pressure by
pulsing a control signal to
a switch to intermittently apply a reverse air pressure to slow the piston
down to match the target
braking speed. In a similar manner, the magnetic rotary brake can be applied
in a continuous
mode or can be applied in an intermittent mode where the brake is pulsed off
and on to slow the
piston down to match the target braking speed. Once the actuator position is
in a final section at
X(tZ) close to the target position X(tf), an output stop signal is sent to the
magnetic brake (B) to
stop the piston. If the actuator is actually stopped at the target position
X(tf), then air pressure is
made neutral on both sides of the piston. Otherwise, forward air pressure is
reapplied to the
piston and the entire braking process is repeated. By attempting to control
the speed of the
piston to match the calculated braking speed, this system seeks to regulate
the pneumatic
actuator in a way that can tolerate and compensate for changes in the system,
including changes
in the brake. In practice, however, the combined use of both the mechanical
braking
arrangement and the air braking arrangement simply increases the number of
variables that must
be programmed into and handled by the control system, thereby complicating the
control
arrangement rather than simplifying it. Again, the dashed lines represent
expected real world
behavior under varying conditions.
Figure 5 is a graph similar to Figure 3 showing the control system and
magnetic
proportional brake arrangement as described in PCT Publ. No. WO 00/53936. Both
a simple and
a sophisticated control system are described for this system. Movement is
initiated at point Xo
by applying a forward differential pressure. The problem with hunting problem
using the simple
control system is best shown in Figure 8 of the PCT application and will not
be discussed. The
sophisticated control system calculates a shut down point XS that is prior to
the desired stopping
point X(tf). This calculation is based upon the kinetic energy of the system
and the available
o~
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
braking force according to the equations as described in the PCT application.
The shut down
point X(ts) represents the exact time that the system will shut down the
differential pressure (P)
applied to the actuator and start the proporational magnetic brake (B). As
shown, the differential
pressure (P) is set to have both the forward and reverse sides of the piston
go to a neutral or
atmospheric pressure at the shut down point X(ts). The control system allows a
user to input a
desired velocity profile for the actuator and uses a low level brake signal to
maintain that
velocity at the programmed desired velocity between X(t,,) and the shut down
point X(ts). In the
event that the proportional magnetic brake does not stop the piston by the
desired location point
X(tf), then the system must release the proportional brake and apply reverse
thrust to move the
piston back toward the desired location point X(tf). While the calculation of
a shut down point
X(ts) based on kinetic energy may be helpful in reducing the hunting problem
of the simple
control system, if there is an overshoot the control system has no alternative
but to reverse thrust
to back the piston up to the desired location point X(tf). Unfortunately, the
control system
responds to an overshoot in such a way that fiuther hunting is likely because
the control system
1 S releases the brake before applying the reverse thrust, thereby creating an
uncontrolled motion
that complicates further positional control.
The control system based on a kinetic energy model also does not address the
real world
problems of precision and repeatability in accurately positioning the piston
at a desired stopping
position. To be effective, the calculation of the shut down point requires
precise knowledge
about the mass of the load and the variability of the braking force. Even if
precise knowledge of
these values can be programmed into the control system, this sophisticated
control system is
unable to accommodate changes in loads during an operation or changes in the
remaining parts
of the system due to friction, wear or variance. Although the challenges
presented by variables
in the differential air pressure are seemingly addressed by setting both the
forward and reverse
sides of the piston to a neutral position, it is believed that this actually
introduces additional
uncontrolled variables into the control process in the form of how the air
pressure on both sides
of the piston actually changes to the neutral or atmospheric pressure. Because
of the extreme
importance of mass in determining the kinetic energy model that is used to
control the system, it
is also questionable as the ability of the system to maintain even this amount
of accuracy in the
event of any significant load change of more than a few percent.
Referring now to Figure 6, a graph showing the operation of a precision servo
control
system for a pneumatic actuator in accordance with the present invention will
now be described.
~3
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
As with the existing pneumatic control systems, the piston is started moving
forward at point
X(to) by applying a forward differential pressure or thrust. Like existing
control systems for
electrical actuators, the precision servo control system of the present
invention allows a user to
input a desired motion profile for the actuator. The precision servo control
system preferably
uses the brake (B) and feedback information from the sensor system to maintain
that motion
profile throughout the acceleration and constant velocity portions of the
motion profile. Unlike
the shut down point X(ts) in the system described in Figure 5, the present
invention continues to
apply a forward thrust and, starting at point X(td),also applies an equal and
opposite backward
thrust. Because thepressures of these two thrusts are equivalent and opposite,
they operate to
effectively cancel out any contribution of the pressurized air in the system
during the
proportional braking period from point X(td) to just prior to the desired
position at point X(tf).
The active application of the same pressurized gas to both the forward and
reverse sides of the
piston guarantees that there will be no complicated interaction of a reduction
of pressures during
the deceleration period.
In addition, the present invention preferably monitors for a predetermined
condition
represented by the point X(tf') just prior to the desired stopping position at
point X(tf). The
predetermined condition represents the point at which the precision servo
control system first
determines that either a defined percentage of a distance from the braking
point to the desired
stopping position has been reached or that there has been a defined percentage
reduction in the
velocity of the piston. Once the point X(tf') has been reached, the reverse
thrust of pressurized
gas applied to the reverse side of the piston is turned off. The predetermined
condition is
selected such that the momentum of the load carried by the piston will be
sufficiently small
enough that it can be accurately controlled by the brake. Removing the reverse
thrust at the point
X(tf') insures that the actuator will have the necessary forward momentum to
arrive at the desired
position at point X(tf) without the need for any jerking, stopping or crawling
of the actuator. In
the event that the momentum of the load carried by the piston does cause an
overshoot, the
precision servo control system locks the brake, then removes the forward
thrust and applies the
reverse thrust, after which the brake is slowly released to return the piston
to the desired stopping
position. In the event that X(tf') is set equal to X(tf), the servo precision
control system will
maintain both forward and reverse thrust until the desired stopping point is
reached.
Unlike the existing control systems that rely purely on air braking through
use of servo
control valves or PWM techniques, the precision servo control system of the
present invention
~ '/
WO 01/31205 CA 02388576 2002-04-25 pCT/L1S00/29877
can use relatively inexpensive directional valves, such as a pair of two-
position valves or a single
three-position valve. More importantly, there is no need for the pressurized
gas supply to be
closely regulated. This means that the precision servo control system of the
present invention
can tolerate contaminated gas lines and other changes over time in the
pressurized gas supply
system because the same pressurized gas is supplied to both sides of the
piston during the
decelaration period. Because of this equal and opposite application of
pressure, the
characteristics of the gas supply on each side of the piston cancel each other
out. As a result, the
precision control system does not need to create complicated higher orders
control functions to
compensate for the very complicated characteristics of a compressible and
changeable gas
supply.
The precision servo control system for a pneumatic actuator of the present
invention
provides a low-cost pneumatic position and velocity control system relative to
electric motion
systems, without the setup and control challenges of traditional pneumatic
servo systems. The
present system is applicable to virtually all types of pneumatic motion
devices including linear
actuators such as rod actuators, rod slide actuators or rodless actuators, as
well as rotary
actuators. The precise control afforded by the present invention provides for
a positional
repeatability of at least +/- 0.1 inches and preferably up to +/- 0.010 inches
independent of the
stroke length of the pneumatic motion device, the pneumatic supply pressure,
or the valve Cv (a
number expressing the ability of a fluid to flow under pressure difference or
pressure drop, also
referred to as flow capacity or flow coefficient). As a result, a user of the
system is able to
define a motion profile for the pneumatic motion device by defining a motion
profile in a manner
consistent with that of servo systems for electrical actuators.
I. System Components
While the system of the present invention is applicable to virtually all
pneumatic motion
devices, for clarity the system will be described below with reference to a
specific pneumatic
motion device, i.e., a linear actuator. With reference to Figure 7, system 1
of the present
invention generally comprises the linear actuator 2, with an integral brake 52
and sensor system,
preferably an encoder 44, and a programmable controller/drive 100 that is in
electrical
communication with the brake 52 and encoder 44. System 1 with linear actuator
2 may be used
in both horizontal and vertical applications.
/J
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
LA. Linear Actuator -- Rodless Cinder
The linear actuator 2 has a support member and a driven member moveable along
such
support member. The driven member includes mechanism that connects to a load.
One such
linear actuator or motion device is commonly referred to as a rodless or
pressure cylinder. In
general, a rodless cylinder comprises an elongated cylinder body and an
internal bore and an
elongated slot extending through the cylinder body wall. A piston is
reciprocally moveable
within the bore in response to the introduction and exhaust of pneumatic
pressure on opposite
sides of the piston and a carrier or transfer bracket is connected to and
moveable with the piston
for connection to a load. Various embodiments of these types of rodless
cylinders exist in the art.
Examples of such rodless cylinders are illustrated in U.S. Patent No.
4,545,290 and U.S. Patent
No. 5,555,789 both of which are incorporated herein by reference. The general
structure of such
a rodless cylinder is illustrated in Figure 7.
With reference to Figures 7 and 8, the rodless cylinder in accordance with a
preferred
embodiment of the present invention includes an elongated cylinder body 2
having an internal
1 S bore 3 and an elongated slot 4 extending through the wall of the cylinder
body 2. Positioned at
opposite ends of the cylinder body 2 are sealing heads 22 and 24 to define a
pair of pressure
chambers 7 and 8. A reciprocally moveable piston S is positioned within the
bore 3 and is
adapted for reciprocal movement back and forth within the bore 3 in response
to fluid pressure
introduced into or exhausted from the fluid pressure chambers 7 and 8.
Specifically, the fluid
pressure chamber 7 is defined between the piston 5 and the sealing head 22,
while the fluid
pressure chamber 8 is defined between the piston 5 and the sealing head 24. A
carrier or transfer
bracket 6 is connected with the piston 5 and extends through the cylinder slot
4 where it is
connected with a load (not shown). Sealing means in the form of an elongated
sealing band 90
(Figure 8) is provided to seal the slot 4, and thus the pressure chambers 7
and 8, on opposite
sides of the piston 5 as the piston 5 moves within the bore 3. The sealing
band 90 is retained in
the sealing heads 22 and 24 by a retaining member 91 and one or more set
screws 92. A
magnetic dust band 93 can, if desired, also be provided to seal the top of the
slot 4. The dust
band 93 is retained in the sealing heads 22 and 24 by a retaining member 94
and the set screws.
The details of such sealing means and dust cover as well as the detailed
structure of the rodless
cylinder as described above is well known in the art as exemplified by U.S.
Patent, Nos.
4,545,290 and 5,555,789. Although the pressure used to drive the piston 5
within the cylinder
bore 3 can be any type of fluid pressure, it is preferably pneumatic or air
pressure.
/~
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
Figure 9 is a diagram of a system 10 including an exploded view of a
positioning
assembly 11 usable with the rodless cylinder as described above for
controlling the movement
and positioning of the piston 5 (Figure 7) and thus the load.
The assembly 11 includes an elongated flexible member shown in the preferred
embodiment as a belt 12. The belt 12 is typically constructed of a rubber or
synthetic material
and may include a ridged or toothed inner surface as shown by the ridges 13.
Although the
member 12 is flexible to permit it to travel around-the pulleys as described
below, it is also
preferably non-extendable. In the preferred embodiment, a pair of belt clamp
portions 18 and 20
are attached to one end of belt 12 through bolts or other types of fasteners
and a pair of belt
clamp portions 14 and 16 are attached to the other end of belt 12 through
bolts or other types of
fasteners. The belt clamps 16 and 20 include connector portions 15 and 19,
respectively, which
fasten to opposite ends of the Garner or transfer bracket 6 (Figure 1).
Preferably, the connector
portions 15 and 19 include threaded apertures 15a and 19a for receiving
threaded screws to
connect the belt clamps and the ends of the belt 12 to the Garner bracket. A
carrier or transfer
bracket includes any structure or apparatus moveable with the piston for
connecting the
elongated flexible member or belt with the piston.
Other types of apparatus or structure for connecting the elongated flexible
member or
belt 12 with the transfer bracket or piston may also be used. For example, the
belt 12 may be an
endless belt which is simply clamped or otherwise connected with the transfer
bracket. The
flexible member may also comprise an endless cable connected to the transfer
bracket or a cable
with ends connected to the transfer bracket, similar to the belt 12.
The elongated flexible member or belt 12 is connected with a motion conversion
device
for converting linear movement of the piston into rotary movement. In the
preferred
embodiment, such device comprises a pair of pulleys 30 and 32 around which the
belt 12 travels.
Each pulley 30 and 32 preferably includes a plurality of ridges or teeth 31
and 33 for mating with
the toothed inner surface of the belt 12. The pulleys 30 and 32 are mounted at
opposite ends of
the rodless cylinder as shown in Figures 7 and 8. In particular, the pulley 32
is rotatably
mounted on a shaft 45 which is in turn attached to a mounting bracket 50
within a pair of
apertures SOb and SOc. The mount bracket 50 is generally a U-shaped bracket
having a base SOa
attached to an inner surface portion 48d of a pulley housing 48 using bolts or
other fasteners. A
pair of mounting plates 34 and 36 are attached to opposing sides 48b and 48c,
respectively, of
the housing 48 using bolts or other fasteners.
7
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
The pulley 30 is rotatably mounted on a rotation shaft 42 which is supported
within the
housing 46. A pair of mounting plates 38 and 40 attach to opposing sides 46b
and 46c,
respectively, of the housing 46. As shown, the mounting plates 38 and 40
include apertures 38a
and 40a for rotatably supporting the shaft 42. One end of shaft 42 is
connected with a
proportional brake 52 through the aperture 38a, while the other end of shaft
42 includes a portion
43 connected to a rotary encoder 44 through the aperture 40a. The apertures
38a and 40a may
each include bearings for accommodating rotation of the shaft 42.
Although the pulley 32 may rotate relative to the shaft 45, if desired, the
pulley 30 is
connected with its shaft 42 for rotation therewith. Thus, rotation of the
pulley 30 causes
corresponding rotation of the shaft 42. The belt 12 extends around the pulleys
30 and 32 so that
rotation of the pulleys causes movement of the belt or piston and movement of
the belt or piston
causes rotation of the pulleys.
The brake 52 is connected to the shaft 42 for rotation therewith and
preferably may be
any brake whose braking force is proportional to an input signal or driving
force. Alternatively,
a non-proportional brake could be utilized with a pulse width modulation
control or similar
intermittent control technique that can selectively exert a braking force. In
the preferred
embodiment, the brake 52 is an electrical current proportional brake whose
braking force is
proportional to an electric current input. More specifically, an example of a
current proportional
brake usable in the present invention is a Lord Rheonetic magnetic particle
brake sold the by
Lord Corporation. In the preferred embodiment, the shaft 42 is connected with
the brake 52 as
an output shaft. Thus, braking forces from the brake 52 are applied to the
shaft 42.
The encoder 44 may be any type of encoder capable of providing an output
signal
indicating the relative position of a piston in a rodless cylinder. In the
preferred embodiment, the
encoder 44 is a rotary encoder connected with the shaft 42 and functioning to
provide a signal
indicating the relative angular position of the shaft 42 and thus the linear
position of the piston 5
(Figure 7). An example of a rotary encoder usable in the present invention is
a Hewlett-Packard
HP HEDS 5500 rotary encoder used with an AMP encoder cable or a Renco RM15
series
encoder. Linear encoders may also be used in place of the rotary encoder
wherein the linear
encoder provides true carrier 6 position instead of the position as determined
through movement
of the belt 12. Alternatively, other types of sensor systems for detecting
position, velocity or
acceleration may be utilized with the present invention, such as optical,
infrared, photoelectric,
laser, microwave, accelerometers or the like.
/8
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
The pair of opposed sealing heads 22 and 24 are connected to the ends of the
cylinder
body for sealing the bore 4 in the cylinder 2 and defining the chambers 7 and
8. The front face
48a of the housing 48 is connected to the sealing head 22 through bolts or
other fasteners. The
sealing heads 22 and 24 as well as the cylinder body 2 may be made from a die-
cast aluminum or
other material. As shown best in Figure 2, the sealing heads 22 and 24 include
a recessed area
96 between the seal member 90 and the dust band 93 to accommodate the belt 12.
As shown best in Figs. 7-11, a belt containment portion 26 is attached to the
bottom of
the cylinder body 2 by a plurality of threaded connectors extending through
the holes 27. The
portion 26, which may be integral to the extrusion of the cylinder body 2,
includes a belt channel
28 for accommodating the belt 12. In operation, the belt 12 moves freely
within the channel 28
between the portion 26 and the bottom of the surface of the cylinder bottom.
Preferably the
portion is connected to the cylinder 2 on a side opposite the slot 4 and
opposite the Garner
bracket.
The sealing heads 22 and 24 include ports 70 and 72, respectively, for
receiving
pressurized air or other fluid from a pressurized air source 54, and also
preferably include an air
cushion port 21 for cushioning the movement of the piston 5 as it approaches
the end of the
cylinder. The air source 54 includes two lines 66 and 68 for providing
pressurized air to the
valves 74 and 76, and ultimately to the ports 70 and 72 through the lines 67
and 69. The air
source 54 is preferably a constant pressure source and valves 74 and 76 are
preferably standard
directional air valves (no servo-valves are required). The air pressure range
is preferably in the
to 120 psi range and, more preferably, in the 50 to 100 psi range.
In a first preferred embodiment, valves 74 and 76 comprise two 2-position 3-
way
normally open solenoid-actuated valves, as shown in Figure 12. In a second
preferred
embodiment, valves 74 and 76 are unitary in the form of a 3-position 4-way
spring centered
25 dual-solenoid valve, as shown in Figure 13. The solenoid valves are
preferably selected so as to
have a response time of less than 20 milliseconds to minimize the effect the
response time has on
system 1. The valves 74 and 76 are controlled directly by controller/drive 100
and are preferably
mounted in close proximity to cylinder 2; short air line lengths between the
valves 74, 76 and
cylinder 2 help to optimize the response time of system 1.
For horizontal applications of system 1 with rodless cylinder 2, either the
valve
configuration of Figure 12 or the valve configuration of Figure 13 may be
used. When plumbing
the rodless cylinder 2, keeping the valves 74 and 76 as close to the cylinder
2 as possible and
l9
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
using air lines of the same length will help to optimize the performance of
system 1. Pressurized
air source 54 is preferably operated at the minimum supply pressure necessary
for the application
at hand. This helps to minimize the heat generated by the brake 52, therefore,
helping to
maximize the application duty cycle.
For vertical applications of system 1, a dual pressure system is preferably
used to obtain
optimum performance. The dual pressure system may be achieved by using two
separate
external pressure regulators or a valve with a sub-base mounted pressure
regulator. Further,
vertical applications are preferably connected by using two separate valves to
provide the two
separate pressures to rodless cylinder 2 (similar to Figure 12). Moreover, the
brake end of
rodless cylinder 2 is preferably mounted at the top in the vertical
application to help minimize
the chance of slack in the belt 12 at the brake 52. Again, as with horizontal
applications, the
valves 74 and 76 are preferably mounted as close as possible to cylinder 2 and
air lines of equal
length are preferably used to help optimize the performance of the system 1.
However, different
from the horizontal applications, air source 54 is preferably operated at
different pressures for
each direction of movement to compensate for gravity's effect on the load and
to help minimize
the amount of brake torque needed.
As shown best in Figure 9, the valves 74 and 76 in the lines 66 and 68 control
the
introduction of pressurized air into and exhaustion of air from the ports 70
and 72 and thus the
chambers 7 and 8 (Figure 7). For example, by opening the valve 76, while the
valve 74 remains
closed, air pressure is provided to the port 72 (and thus the chamber 8) and
exhausted from the
port 70 (thus the chamber 7) through the closed valve 74. This causes the
piston 5 to move away
from sealing head 24 (toward the right as viewed in Figure 7). Likewise, with
the valve 74 open
and the valve 76 closed, air pressure is provided to the port 70 and exhausted
from the port 72.
This causes the piston to move away from the sealing head 22 (toward the left
as viewed in
Figure 7). With both valves 74 and 76 open, the piston 5 may be held in place
by equalizing the
pressure in the chambers 7 and 8 on opposite sides of the piston.
LB. Programmable ControllerlDrive
The programmable controller/drive 100, a block diagram of which is depicted in
Figure
14, generally comprises a processor 102 in communication with a memory storage
device 104,
and input/output (I/O) interface 106, a display 108, a user-input device 110,
and a
communication port 112. The processor 102 operates to control the operation of
system 1 by
~D
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
executing pre-programmed functions and programs) stored in the memory storage
device 104.
In a preferred embodiment, the processor 102 is a digital signal processor
(DSP), such as the
Texas Instrument TMS320C240. The memory storage device 104 preferably
comprises an
EEPROM that is capable of storing demo programs and up to 7 - 256 line
programs that can be
activated directly, on power-up of controller drive 100, or with an input. The
I/O interface 106
preferably provides for electrical connection to each of the solenoids which
direct the operation
of the valves 74 and 76 (along lines 80 and 82, see Figure 9) as well as
electrical connection to
brake 52 (along line 84, see Figure 9) and to encoder 44 (along line 78, see
Figure 9). Display
108 is preferably in the form of an LCD screen that is unitary with the
controller/drive 100 while
the user-input device 110 is preferably in the form of a keypad that is also
unitary with the
controller/drive 100. The communication port 112 is preferably an RS-232 port
enabling
communication of controller/drive with an external PC 114 for external
programming and data
collection. Controller/drive 100 is preferably powered from line voltage and
therefore does not
require an external power supply. A preferred facial configuration of
controller/drive 100 is
shown in Figure 15 with the keypad 110 and LCD 108 screen noted.
I. C. Alternative System Components
As indicated above, the system 1 may be modified as appropriate to a specific
application
without departing from the spirit or scope of the invention. By way of non-
limiting example,
these modifications may include: (1) a rodless cylinder that utilizes a linear
encoder 115 placed
along the body 2 of the cylinder, as opposed to the rotary encoder 44, wherein
the read head 116
of the encoder 115 is attached to the carrier 6, see Fig.l6; (2) a rodless
cylinder that utilizes
linear, friction pad braking devices 117 on the carrier 6 that are driven by
solenoids 118, see Fig.
17; (3) a slide rod actuator 119, see Fig. 18; (4) a rotary actuator 121
utilizing a vane and a stop,
see Fig.l9.
II. System Operation
In the system comprising a pneumatic motion device, such as the rodless
cylinder 2
described above, and the programmable controller/drive 100, the programmable
controller/drive
100 operates to control the position and velocity of rodless cylinder 2.
Specifically,
programmable controller drive receives signals from encoder 44 that reflect
the linear position of
the piston 5 in the rodless cylinder 2. The position signals are processed
based upon a stored
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
motion profile to generate control signals. The controller/drive outputs the
control signals to the
valves 74 and 76 controlling the introduction and exhaust of pressurized air
into and out of the
pressure chambers 7 and 8 and to the proportional brake 52 for use in
controlling movement of
the piston 5. By operating on the position signals from the encoder 44, the
controller/drive 100
effectively uses feedback information to precisely control movement of the
piston S at desired
speeds and to the desired positions. The ability to control such movement of
the piston 5 permits
the use of the system 1 for a wide variety of applications. System 1 may be
used for traditional
applications, as well as various applications requiring precise and repeatable
movement
including but not limited to, picking up a load at one position and releasing
it at another,
commonly referred to as "pick and place."
II. A. System Component Functions
In general, the operation of system 1 may be described as follows, the piston
S in the
cylinder bore 3 is caused to move reciprocally as a result of the selective
introduction and
exhaustion of air pressure into and out of the pressure chambers 7 and 8.
Because the piston is
connected with the carrier bracket 6 which is in turn connected to a load,
such movement of the
piston 5 results in corresponding movement of the load. Also, because the belt
12 is connected
at its ends to the bracket 6, movement of the piston results in corresponding
movement of the
belt 12 around the pulleys 30 and 32. Accordingly, the belt 12 functions to
convert the
substantially linear movement of the piston 5 to rotary movement of the
pulleys 30 and 32, and
in particular the pulley 30. By use of the rotary encoder 44 connected to the
shaft 42 of the
pulley 30, the angular position of the shaft 42 can be monitored by processor
102. Because the
angular position of the shaft 42 is directly related to the linear position of
the piston 5 by virtue
of the belt 12, the rotary encoder 44 can be utilized to monitor the movement
and positioning of
the piston 5 and thus the connected load.
Specific operation of system 1 is defined by a user-programmed motion profile
that is
entered through the user-input 110 or external PC 114 and stored in the memory
storage device
104. The motion profile is defined by a programmed move distance, a maximum
speed,
acceleration time, and deceleration time for piston 5 within cylinder bore 3.
The programmed
move distance specifies, for example, either an absolute or an incremental
distance. The
absolute distance specifies a distance from a home position, which is the
piston S stopped at
either end of the cylinder against sealing heads 22 or 24, while the
incremental distance specifies
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
a distance from a current position of the piston 5 to a new position. Home
position is preferably
determined automatically by the controller/drive 100 upon power up or upon
receiving a home
instruction and is achieved by moving the piston 5 slowly until it reaches an
end of the cylinder
or a hard stop. Homing is performed while gas is applied for forward or
backward motion of the
piston 5 and while the brake 52 is engaged to maintain the piston 5 at a slow
homing speed.
Upon reaching the home position, the encoder counter is set to zero.
The maximum speed is the maximum desired speed that the piston 5 achieves
during
acceleration. The acceleration time specifies the amount of desired time for
the piston 5 to
accelerate from zero (no movement) to the entered maximum speed. The
deceleration time
specifies the amount of desired time for the piston S to decelerate from the
maximum speed to
zero (no movement). A position tolerance may also be entered by the user to
define the motion
profile, or a default position tolerance may be used. The position tolerance
specifies the amount
of acceptable error between the desired and actual end position of the piston
5. For example, it
may specify a repeatability of up to 0.01 inches such that the actual position
of the piston 5 is
within 0.01 inches of the desired position on successive movements. A holding
torque may
further be entered by the user, or a default value accepted. The in-position
holding torque may
be set from 0 to 100% of the maximum brake torque of the brake 52. During
dwell time (non-
motion of the piston 5), especially in horizontal applications, it is often
possible to reduce the
torque necessary to hold a load. By reducing the holding torque, as in stepper
systems, the
efficiency is improved and the holding device is not required to dissipate as
much heat, therefore,
improving the allowable duty cycle.
Based upon the stored motion profile, various graphical examples of which are
shown in
Figure 20, the processor 102 executes the program stored in memory storage
device 104 to
receive position signals from the rotary encoder 44, process the position
signals, and provide
appropriate control signals to the valves 74 and 76, and to the brake 52. The
controller/drive
100, through the processor 102, utilizes feedback techniques (described in
further detail below)
to control the movement of the piston 5 substantially consistent with the
programmed motion
profile. As shown in Figure 20, the motion profiles established by the user
are generally of a
trapezoidal configuration, wherein the left-hand side indicates the
acceleration of the piston 5,
the top flat portion indicates a constant velocity of the piston 5, and the
right-hand side indicates
the deceleration of the piston 5. It should be noted that the motion profile
may differ from a
trapezoidal configuration in response to programmed position parameters.
:~ 3
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
More particularly, the controller/drive 100 provides control signals on the
lines 80 and
82 to control opening and closing of the valves 74 and 76. It also provides a
control signal on
the line 84 to control the brake 52. With the proportional brake 52 of the
preferred embodiment,
the control signal on the line 84 controls the magnitude of the current
provided to the brake 52
and thus the amount of force by the brake 52 applied to shaft 42. By
controlling the magnitude
of the braking force applied to the shaft 42 in combination with the air
pressure provided to or
exhausted from the chambers 7 and 8, the movement of the piston 5 can be
controlled. Such
control can include the position at which the piston is stopped as well as the
speed or motion
profile at which the piston moves to the desired stop position. The control
signals are generated
through processing of the position signals received from the rotary encoder
44. In particular, the
rotary encoder 44 provides signals indicating a relative angular position of
shaft 42, thus,
providing an indication of a corresponding linear position and movement of
belt 12 as it rotates
the shaft 42 via the pulley 30.
ILB. Control Scheme
Figure 21 is a flow chart that provides a detailed view into the operation of
system 1, and
specifically the control loops utilized within controller/drive 100 to achieve
the operation
described above. Upon receiving the motion profile parameters, i.e., move
distance, maximum
speed, acceleration time, and deceleration time, and upon receiving a position
command from a
user or program, per input block 120, controller/drive 100 operates to
generate the motion profile
based on the specified acceleration time, deceleration time and speed (and any
motion profile
adjustments determined in a previous cycle), per operations block 122. In
doing so, the
controller/drive 100, having previously established the home position and the
encoder count
corresponding to the home position, converts the move distance and the
position command into
corresponding encoder counts. Therefore, as the controller/drive 100 detects
position signals
from the rotary encoder 44, which it receives as counts indicating a relative
angular position of
shaft 42, it can readily compare the detected encoder counts with the encoder
counts calculated
from the motion profile information.
Next, processor 102 of controller/drive operates to compare the current
position of the
piston 5 with the commanded position of the piston ~, per decision block 124.
In particular, the
present position is detected through a signal on the line 78 received from
rotary encoder 44. That
detected present position is compared with the motion profile to determine if
the position error
~y
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
with respect to the commanded position is less than the position tolerance. If
the current position
is within the position tolerance, the piston 5 is at the commanded position,
the move is complete,
and the valves 74 and 76 are opened to equalize the pressure on either side of
the piston 5.
Controller/drive 100 then awaits a new position command, per input block 120.
If the current position is not within the position tolerance, the
controller/drive 100
operates to store the amount of position overshoot or undershoot, per
operations block 127.
Preferably the overshoot or undershoot values are placed into a32x32 array
based on 32 actuator
regions that span the stroke of the actuator. In the situation of an overshoot
of the target position
by the piston 5, the starting region, ending region and the amount of
overshoot in distance are
noted in the array. In the case of another move in the same starting and
ending regions, per input
block 120, the overshoot distance will be compensated for by beginning the
deceleration at an
earlier position, i.e. the amount of overshoot distance is subtracted from the
beginning
deceleration position. In the situation of an undershoot, i.e., the commanded
motion profile
indicates that the target position should have been reached at this time but
has not yet been
1 S reached, the starting region, ending region, and the amount of undershoot
(the calculated distance
from the target position) are noted in the array. In the case of another move
in the same starting
and ending region, per input block 120, the undershoot distance will be
compensated for by
delaying the deceleration by the undershoot distance. The overshoot/undershoot
adjustment is
calculated per operations block 129.
After storing any overshoot or undershoot values, the controller/drive 100
next
determines whether the current position of piston 5 is the beginning
deceleration position as
determined by the motion profile, per decision block 126. If the current
position is indeed the
beginning deceleration position, a deceleration constant, KT is integrated
into the control loop,
per operations block 128 and the valves 74 and 76 are switched by
controller/drive 100 to force
pressurized air to act on both sides of the piston 5. As described in
connection with Figure 6,
this equal and opposite pressurization of both the forward and reverse sides
of the piston 5 aids
in the deceleration of the piston 5 and limits the load on the brake 52 and
the belt 12. By
adjusting valve Cv (a number expressing the ability of a fluid to flow under
pressure difference or
pressure drop, also referred to as flow capacity or flow coefficient), the air
recirculation can
significantly help to stop the load or load atop the piston 5. Once the piston
5 reaches a
predetermined deceleration distance or speed, preferably 75% of the
deceleration distance, or
25% of the constant speed, whichever comes first, the valves 74 and 76 are
switched back to the
~J
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
same position as prior to starting the deceleration, per operations block 132
(the changing of the
valves is noted in the motion profile of Figure 23). Note that KT will
continue to influence the
deceleration until the piston 5 is completely stopped, as can be seen in
Figure 22.
If the current position of the piston S is not the beginning deceleration
position, the
controller/drive 100 controls movement of the piston ~ to perform the
commanded move. In
doing so, the controller/drive 100 determines a motion direction through the
signal from the
rotary encoder 44 and opens the valve 76 or the valve 74 depending on the
direction of motion
determined by the motion profile, per operation block 134. The
controller/drive 100 then
monitors the position and velocity of the piston 5 through the processing of
position signals
received from the rotary encoder 44, per operations block 136.
Having converted the programmed command speed and position information into
encoder counts, the controller/drive 100 is able to compare them to the
monitored speed and
position, which are provided as encoder counts from the rotary encoder 44 to
the controller/drive
100. Specifically, controller/drive 100 operates to compare the current speed,
as determined by
the rate of encoder counts from the rotary encoder 44, with the programmed
command speed
according to the motion profile, per operations block 138.
If the speed or velocity of piston 5 is too fast compared to the command
speed, as
determined per decision block 140, the current to the brake 52 is increased,
per operations block
142, thereby increasing the braking force, slowing the rotation of the shaft
42, and slowing the
linear movement of the piston 5. If the speed or velocity of the piston 5 is
not too fast, the
controller/drive 100 determines if the speed or velocity of piston 5 is too
slow compared to the
command speed, per decision block 144. If the speed of the piston 5 is too
slow, the current to
the brake 52 is decreased, per operations block 146, thereby decreasing the
braking force,
speeding the rotation of the shaft 42, and speeding the linear movement of the
piston 5.
In controlling the brake 52, the controller/drive 100 also compares a current
signal with
the commanded current, per operations block 148. The current signal is the
amount of electric
current to the brake 52. The commanded current is calculated according to the
following
equation:
(1) KP*~x+KV *4v
where: KP = proportional gain, a position constant
dx = the position error from the motion profile
KV = derivative gain, a velocity constant
WO 01/31205 CA 02388576 2002-04-25 pCT/jJS00/29877
0v= a velocity error calculated from a difference between the actual speed of
the
piston 5 as determined by a rate of the received encoder counts and the speed
in the
motion profile.
If the comparison indicates that the signal current is within an acceptable
error threshold
of the command current, as determined per decision block 150, the move of the
piston 5 is
complete and the controller/drive 100 returns to the monitoring of the
position and velocity of
piston, per operations block 136. If the comparison indicates that the signal
current is outside the
acceptable error threshold, the controller/drive 100 executes a current loop
to determine an
amount by which to increase or decrease the current to the brake 52.
If the signal current is below the calculated command current, as determined
per decision
block 152, the controller/drive 100 operates to increase the current to the
brake 52, per
operations block 154. If the signal current is above the calculated command
current, as
determined per decision block 156, the controller/drive 100 operates to
decrease the current to
the brake 52, per operations block 158. The controller/drive 100 then operates
to loop back to
operations block 148 wherein the current signal is once again compared by the
controller/drive
100 against the command current for the determination of whether the signal
current is within an
acceptable error threshold and, correspondingly, whether the move of the
piston 5 is complete.
Figure 22 describes the control scheme of the system 1 in a standard block-
diagram
format. The control loops shown within the scheme include: ( 1 ) a position
loop 170 for feedback
comparison of the commanded position and the actual position of the piston 5,
which generally
comprises the blocks of 120, 122, and 124 of the flowchart of Figure 21; (2) a
velocity loop 172
for feedback comparison of the commanded velocity and the actual velocity of
the piston 5,
which generally comprises the blocks of 134 through 146 of the flowchart of
Figure 21; and (3) a
current loop 174 for feedback comparison of the commanded brake current and
the actual brake
current supplied to the brake 52, which generally comprises the blocks of 148
through 158 of the
flowchart of Figure 21.
The above three loops utilize the standard control parameters of a servo-
system
including: (1) proportional gain, KP; (2) integral gain, KI; and (3)
derivative gain, KV. The
proportional gain, KP, is the position error gain that determines how
sensitively the
controller/drive 100 will respond to the position error (difference in command
and actual
position). The controller/drive 100 responds to position error more
effectively when increasing
the KP gain. However, setting the KP gain too high will lead the system 1
toward positioning
WO 01/31205 CA 02388576 2002-04-25 pCT/US00/29877
instability. As such, the controller/drive 100 is preferably programmed with a
user-changeable
default gain on the order of 24.
The integral gain, KI, is the position error integral gain. The
controller/drive 100
operates to accumulate position error while the piston 5 is approaching the
target position and
multiply the error by the integral gain. The result is used by the
controller/drive 100 to
determine the required brake current for positioning. The controller/drive 100
is preferably
programmed with a user-changeable default integral gain on the order of 1.
The derivative gain, KV, is the speed error gain. The KV gain determines how
effectively the controller/drive 100 responds to the speed error while the
piston 5 is in motion.
Increasing the KV gain reduces the speed following error. However, jerky
motion occurs if the
KV setting is too high. As such, the controller/drive 100 is preferably
programmed with a user-
changeable default derivative gain on the order of 4.
In addition to the feedback loops, the block diagram depicts a deceleration
compensation
path 176 within system 1 that enters into effect upon the piston 5 beginning
its deceleration, as
determined by the motion profile. The deceleration compensation path 176
generally comprises
the blocks of 126 and 128 of the flowchart of Figure 21. The deceleration
compensation path
176 uses a control parameter beyond the standard PID parameters of a servo-
system. This
control parameter is a deceleration current constant gain, KT. The KT gain is
used to set the
minimum brake current while decelerating the load that is secured to the
carrier 6. It is used to
adjust for position overshoot, position undershoot, and deceleration profile
linearity when the
carrier 6 approaches the final position. The controller/drive 100 is
preferably programmed with a
user-changeable default KT gain on the order of 48. The deceleration
compensation path 176 is
also the path that initiates the energizing of both of the valves 74 and 76 to
aid in the braking of
the piston 5, as indicated by path 178 (which is also reflected in block 130
of the flowchart of
Figure 21 ).
The four gain parameters KP, KI, KV, and KT allow the controller/drive 100 to
compensate for varying loads, variations in gas supply pressure (e.g., system
1 is able to
maintain a positional profile in view of pressure variations), and temperature
variations.
Specifically, the gain parameters and the controller/drive 100 allow system 1
to have a position
repeatability (tolerance) of +/- 0.01 inches, or absolute value of 0.01,
regardless of stroke length.
Stroke length is preferably in the range of 24 inches to up to SO feet or more
depending upon the
capability of design the actuator to support such stroke lengths. Further,
system 1 is able to
WO 01/31205 CA 02388576 2002-04-25 pCT/LJS00/29877
handle changes in load without any changes in the gain parameters. For
example, the piston S
and carrier 6 may carry a first load from a first position to a second
position and a second load
from the second position to a third position, wherein the first and second
load vary by at least
33% (up to the entire range of the load specification of the actuator and
brake). In doing so,
system 1 is able to maintain its position repeatability of +/- 0.01 inches
without any change in the
gain parameters. Note that the third and first position may be the same or
different positions. A
portion of path 178 overlaps that of path 180, which is the output path from
the position control
loop 170, whereby any motion of the piston 5 that is required as determined by
the position
control loop 170 is initiated by determining the direction of the motion
(sign) and activating the
appropriate forward or backward valve 74 or 76, corresponding to block 134 of
the flowchart of
Figure 21. The actual position of the piston 5 is calculated by the
controller/drive 100 by
initially summing the forces acting on the piston 5 at summing junction 182.
These forces
include: (1) a thrust force, which is determined from the pressure difference
between chambers 7
and 8; (2) a stopping force, which is determined from the brake torque and
pulley pitch radius;
and (3) an external friction force, due to the friction between the piston S
and the cylinder bore 3.
The result of the summed forces is the net force necessary to drive the load
secured to carrier 6 to
the command position.
The net force creates the motion that is detected by the controller/drive 100
and from that
motion the controller/drive 100 detects the position of the piston 5. Velocity
is further derived
from the position signal. Upon reaching the command position, within the
desire tolerance, any
overshoot or undershoot in the position of the piston 5 is recorded and
utilized for future
compensations to the KT gain, per path 184, which corresponds to blocks 127
and 129 of the
flowchart of Figure 21.
The operation of system 1 may also be understood with reference to the motion
profile of
Figure 23. As indicated above motion profiles are user-defined by programming
the move
distance, the maximum speed, the acceleration time, and the deceleration time.
How well the
actual motion of system 1 follows the theoretical profile is based on the
system supply pressure,
system C,,, the load, and how well system 1 is tuned. An appropriately tuned
system can
generally assure a maximum velocity and deceleration to be within 10% of the
programmed
values. During acceleration, indicated by the item number 200 on Figure 23,
the current is
applied to the brake 52 as required to limit the rate of acceleration to the
programmed value. The
maximum ramp rate of the acceleration is limited by traditional pneumatic
cylinder limitations.
WO 01/31205 CA 02388576 2002-04-25 pCT/jJS00/29877
During the constant velocity portion of the profile, indicated by item number
202 on Figure 23,
the information from rotary encoder 44 is evaluated to supply a brake current
necessary to
maintain the maximum speed.
Deceleration, indicated by item number 204, is the most tuning critical
portion of the
profile. In order to reduce the high speed thrust requirement of the belt 12
and torque of the
brake 52, both of the valves 74 and 76 are de-energized to allow air
circulation through the
valves 74 and 76 during the initial part of the deceleration. The gain
settings controls how linear
the deceleration is over the programmed deceleration time. If the ratio of KT
to KI is not
appropriately set, a rapid deceleration will be experienced with a short slow
move into final
position. However, with a properly tuned system, the positional repeatability
of the system 1 can
be programmed to +/- 0.010 inches (+/- 0.254 mm) with increments of 0.006
inches (0.15 mm).
ILC. Auto Tune
The controller/drive 100 preferably incorporates a digital auto tuning program
to enable
easy user setup of system 1. All automatic adjustments are preferably made
within the software
that sets the system gain parameters, thus eliminating the time-consuming
adjustments required
by potentiometers. The auto tuning program implemented in the controller/drive
100 firmware
cycles the carrier 6 to positions between 1/4 to 3/4 of the entire stroke
length while searching for
an optimal set of servo parameters. The user must, however, specify the bore
size and usable
stroke length of system 1 before conducting any auto tuning. The auto tune
routine preferably
moves the carrier 6 up to 15 cycles while finding the best set of servo-
parameters, and can do so
with or without a load. In determining the best set of parameters, the auto
tune function of the
controller/drive 100 looks to three factors: (1) overshoot of the position of
the piston 5; (2)
undershoot of the position of the piston S; and (3) velocity following error
(overshoot/undershoot). The goal of the auto tune function is to minimize
these factors and does
so by adjusting the various gain parameters that are set as the defaults
within the firmware of the
controller/drive 100, e.g., adjusting KT for overshoot, adjusting KV and KI
for undershoot, and
adjusting KV for velocity following error. KV, KI, and KT may changed at the
same time or on
an individual basis. The proportional gain KP is not adjusted during auto
tune.
30
CA 02388576 2002-04-25
WO 01/31205 PCT/US00/29877
III. System Summary
System 1 is designed to be a low cost position and velocity control system
relative to
electric motion systems, without the setup and control challenges of
traditional proportional
valve pneumatic servo systems. System 1 is intended for applications not
requiring positional
repeatability better than +/-0.010 inches. Motion profiles are user defined by
programming the
move distance, the maximum speed, acceleration time, and deceleration time.
System variables,
including supply pressure, valve Cv, the load, and how well the system 1 is
tuned, do not effect
the final repeatability of the system 1, however, they will influence how well
the actual motion
follows the theoretical profile. An appropriately tuned system can generally
assure velocity to be
within 10% of the programmed values, with appropriate air pressure.
System 1 significantly reduces the concerns of changing loads, system
friction, vertical
operation, setup, tuning, instabilities, and long stroke lengths, typically
experienced with
pneumatic servo systems. The approach used in the system 1 accomplishes this
by using the
muscle of air to provide thrust, and a current controlled magnetic particle
brake to provide
proportional braking for position and velocity control. This approach
significantly reduces the
effects of directly attempting to control a compressible fluid through
proportional valves, or
trying to predict when to activate an on/off brake to achieve a desired
position.
The system 1 also considers duty cycle limitations due to heat dissipation
requirements of
the brake 52. Heat generation takes place not only during deceleration of a
load, but also during
regulation at a constant speed. It is desirable to operate at a minimum supply
pressure necessary
for the application. This will minimize the heat generated by the brake 52,
therefore, maximizing
the application's duty cycle. In a vertical application it is recommended to
operate at a different
pressure for each direction of motion, depending on the duty cycle desired.
There may also be a
minimum speed that can be achieved with a given air pressure, determined by
the duty cycle
desired.
Although the preferred embodiment has been described, it will be recognized
that
numerous changes and variations can be made and that the scope of the present
invention is
intended to be defined by the claims.
